Artificial Intelligence (AI) Driven Machine Learning Modeling for Process Characterization of Dynamic Freeze Drying (Lyophilization) After Spray Freezing
Spray Freeze Drying can provide significant improvements to processing in pharmaceuticals and has wide applicability in a diverse array of fields. Yet the process is still relatively new and further understanding will yield additional value.
Recent technological advances in artificial intelligence and machine learning have led to a rise in software platforms that automate many of the model building and visualization tasks. These tools allow scientists to both interpret existing data, make predictions, and identify next best experiments in a cost effective and accessible manner. This accessibility allows trained scientists, even without data science or coding experience, to extract knowledge from datasets and accelerate R&D through targeted experimentation.
In this study, different types of data were assembled from experiments using a wide range of materials and process conditions. A cloud-based platform that combines exploratory data analysis, machine learning modeling, and experiment planning using Bayesian optimization was selected to perform an initial analysis and suggest further work. The results using the Sunthetics platform to derive insight into the dynamic freeze drying step of spray freeze drying processes are discussed.
